Working in knowledge science and analytics for seven years, I’ve created and queried many tables. There are quite a few occasions I ponder, “What does this column imply?” “Why are there two columns with the identical identify in desk A and desk B? Which one ought to I exploit?” “What’s the granularity of this desk?” and many others.
For those who’ve confronted the identical frustration, this text is for you!
On this article, I’ll share 5 rules that can enable you create tables that your colleagues will recognize. Please observe that that is written from the angle of an information scientist. Subsequently, it won’t cowl the standard database design finest practices however concentrate on the methods to make user-friendly tables.
Sustaining a single supply of reality for every key knowledge level or metric is essential for reporting and evaluation. There shouldn’t be any repeated logic in a number of tables.
For comfort, typically we compute the identical metric in a number of tables. for instance, the Gross Merchandise Worth (GMV)
calculation would possibly exist within the buyer desk, month-to-month monetary report desk, service provider desk…